The objectives of this project are: (1) the development of quantitative structure-activity correlations among antitumor agent congeners, (2) the development of general structure-activity relationships among large numbers of unrelated antitumor agents and (3) the use of the correlation data obtained to generate leads for new synthesis and increase the rationale for the testing of new agents. Compound families of current interest are colchicines, purines, 2,4-diaminopyrroloquinazolines, quinoliziniums, benzimidazole mustards, phenanthridiniums, Schiff-base alkylating agents and thiadiazoles.